A VOLUMETRIC ESTIMATION OF PHOSPHORIC ACID.
I099
the oxidation of ferrous in some cases is known’ but this is hardly a case of that sort since nitric acid alone oxidizes ferrous solution in the cold easily, as is seen from the results above. This peculiar result will be thoroughly and carefully investigated. Taking a survey of all of this preliminary work as a whole, i t seems that there is some ground, then, for the following conclusions: I . Dilute nitric acid and dilute mixtures containing hydrogen, nitrate and chloride ions may effect the oxidation of certain substances a t different rates, even when the final products of the two oxidations do not differ in their energy content. 2 . This difference may be due ( a ) to the influence of excess of hydrogen ions where chloride ion is absent, and ( b ) when chloride ion is present, to a possible catalytic action of this ion. 3. The most striking example of chloride ion acting as a catalytic agent is in the oxidation of ferrous salts, when i t acts as a negative catalytic agent.2 Finally i t must again be emphasized that this work here reported is purely preliminary, and serves to outline a field of work in which i t is hoped some important results may be achieved, in the future. NEW YORKCITY.
A VOLUMETRIC ESTIMATION O F PHOSPHORIC ACID. BY J. ROSIN. Received April 26, 1911.
The reaction between silver nitrate and a soluble phosphate whereby silver phosphate is precipitated has served as the basis of several attempts a t a volumetric estimation of phosphoric acid. P e r r ~ twho , ~ was the first to propose a method based on this reaction, precipitated the phosphoric acid with silver nitrate in the presence of acetic acid and determined the residual silver by sodium chloride. Holleman4 converted the phosphoric acid or the mono-alkali salt into the dialkali salt by means of caustic soda and precipitated the phosphoric acid with silver nitrate in the presence of sodium acetate. The residual silver mas then titrated in an aliquot portion by Volhard’s method. Liebermans adds a considerable amount of nitric acid to the solution of the phosphate, then adds an excess of silver nitrate and follows this with ammonia water until blue litmus paper is no longer affected. He then determines residual silver by Volhard’s method. Holleman’s method recommends itself and i t is in fact the best of the Benson, J . Physic. Chenz., 7, 1-14; 7, 356-77 (1903). See also in this connection, Benson (loc. cit.) with respect to the influence of iodide ion, and for other theories. J . SOC.Chem. Ind., 1888,311. 2. anal. Chem., 1895, 85. Analyst, 1909,361.
I IO0
GENERAL, PHYSICAL AND INORGANIC.
methods above cited. However, results obtained by it are far from satisfactory. If a considerable excess of sodium acetate is used, sil\-er acetate is co-precipitated and too high results are obtained. With a moderate escess only of sodium acetate, the acetic acid liberated, although very dilute, dissolves sufficient silver phosphate to yield results from 0.5-1 per cent. too low. It is evident that if phosphoric acid is to be completely precipitated, any free acid present must be neutralized. The products resulting from the interaction of disodium hydrogen phosphate and silver nitrate are frequently represented in the t e s t books by the equation Na,HPO, 3AgN0, = Ag,PO, zNaNO, ”0,. (A) Ostwald’ and Treadwel12express the reaction by the equation 2Na,HPO, + 3.4gN0, = Ag,PO, 3NaN0, NaH,PO,. (B) Neither of these equations is correct. If a solution of sodium phosphate be precipitated with an excess of silver nitrate, the filtrate from the silver phosphate contains from 35-38 per cent. of the phosphoric acid originally used. According to equation (A) very little,: if any, phosphoric acid should be found in the filtrate, while according to equation (B) 50 per cent. of the original phosphoric acid should be found in the filtrate. If the filtrate from the silver phosphate is distilled after the excess of the sil\-er nitrate has been removed by sodium chloride, very little nitric acid will be found in the distillate. In two experiments quantities of sodium phosphate which, according to equation (A) should yield nitric acid equivalent to 92.3 and 85.8 cc. 0 . I Ai caustic soda respectively, were precipitated with an excess of silver nitrate, the excess of the latter removed from the filtrate with a slight excess of sodium chloride, and the filtrate from the silver chloride repeatedly distilled with water into standard alkali, carrying the final distillation almost, but not quite, to dryness. Upon titrating the excess of alkali 3.40 and 2.95 cc. respectively of 0.1Ai alkali were found to have been neutralized by the acid that had distilled over. This acid proved to be nitric acid. These facts show that the reaction will be represented more correctly if expressed by the equation 3Na,HPO, 1 6AgX0, = a.\g,PO, 6NaN0, - H,PO,. ( C ) According to this equation no nitric acid is formed. The small quantities of it obtained in the experiments just described probably resulted from the action of the phosphoric acid formed, on the sodium nitrate, at the end of the distillation when the phosphoric acid became concentrated.
+
+
+
+
+
+
“Principles of General Chemistry.” “Qualitative Analysis.” Enough silver phosphate will, of course, remain dissolved to maintain the eqiiilibrium suggested by the equation. ?
A VOLUMETRIC ESTIMATION OF PHOSPHORIC ACID.
I IO1
I have observed that when a solution of free phosphoric acid containing sodium nitrate is concentrated on the water bath, nitric acid is lost. Equation (C) requires one-third of the original phosphoric acid to be found in the filtrate from the silver phosphate. The presence of more than one-third-35-38 per cent.-is probably due to the solvent action of the liberated phosphoric acid on the silver phosphate. The amount of silver phosphate dissolved is dependent upon the time of contact between it and the phosphoric acid. That the reaction proceeds in the sense indicated by equation (C) was confirmed by the following experiment: 3.0164 grams of Na,HPO, were precipitated with a n excess of silver nitrate and made up to 2 0 2 cc.-allbwing 2 cc. for the volume occupied by the silver phosphate (taking the specific gravity of silver phosphate to lie between 4 and 5) and filtered. From IOO cc. of the filtrate the excess of the silver nitrate was precipitated with a slight excess of sodium chloride, filtered, washed, and filtrate and washings, containing the free acid, divided into two equal parts. One part when titrated with 0.1 iV alkali, using methyl orange as indicator, required 18.0 cc. The other part, with phenolphthalein as indicator, required 35.5 cc. With the latter indicator practically twice as much alkali was required. This is in accordance with the behavior of phosphoric acid towards these indicat0rs.l It is also seen that the quantity of alkali consumed corresponds almost theoretically to the amount required by equation (C)-I 7.7 cc. with methyl orange and 35.4 cc. with phenolphthalein. The fact that the filtrate from the silver phosphate is acid to methyl orange conclusiyely proves that equation (B) is incorrect, since according to this equation there is NaH,PO, formed and this substance is not acid to methyl orange. The experiment just described was repeated several times and the same results obtained-with the phenolphthalein twice as much alkali was required as with methyl orange, and the amount of alkali corresponded very closely t o that required by equation (C). The results of these experiments are presented here. The figures in the first line represent the amount of Na,HPO, used for the quantity titrated. I.
Na,HPO, used, gram.. .................... 0.7541 Cc. 0.1 N KOH used, phenolph. ind.. . . . . . . . 35.5 Cc. 0.1 N KOH used, methyl orange ind.. . . . 18.0 Cc. 0.1N KOH required by equation (C), phenolph. ind ........................... 35.4 Cc. 0.1N KOH required by equation (C), methyl orange ind.. ..................... 17.7 Cc. 0.1 N KOH required by equation (A).. . . 53. I
3.
2.
1.0322
48.5 24.5
4.
0.8691 0.4346 41.5 20.6 20. I
10.2
49.4
40.8
20.4
24.7 74.I
20.4
10.2
61.2
30.6
See Ostwald “Scientific Foundations of Analytical Chemistry,” Engl. edition, p.
128.
II02
GEXERAI,,
PHYSICAL AKD INORGANIC.
While tile method presented in this paper was being worked out, Wilkie’ published a modification of Holleman’s method. He neutralizes the phosphoric acid or its solution in nitric acid with caustic soda, using phenolphthalein indicator, adds an excess of 0.1 -V silver nitrate and then runs in dilute c a u t i c soda till just pink. He then adds 2 cc. 0.1 LV H,SO,, makes up to t-olume and titrates an aliquot portion of the filtrate for residual d y e r by Volhard’s method. -1procedure almost the same as Wilkie’s was tried before IVilkie’s method vias published, but, on account of the precipitation of silver oxide rrvhen the alkali is added, the neutralization is difficult, as Wilkie himself admits, and failing to obtain a sharp end-point, I abandoned further attempts along this line. -1method was sought where the use of bodium acetate could be ax-oided by neutralizing the free acid formed in the interaction of sodium phosphate and silver nitrate with a substance, a moderate excess of which would have no effect upon the excess of silver nitrate. Still under the impression that a t least most of the acid is nitric acid, calcium carbonate a t once suggested itself. It was, however, soon found that the precipitated carbonates of the alkaline earths, and ex-en the naturally occurring carbonates of these earths, limestone, spar, strontianite and witherite, are transposed by silver nitrate solution into silver carbonate and the corresponding nitrate of the alkaline earth. Owing t o their finer state of diiision, th? precipitated carbonates are much more readily transposed than the natural. Precipitated basic zinc carbonate is also transposed, althougli 4owly. Zinc oxide was tried and found to answer the purpose. In a preiiminary series of experiments with disodium hydrogen phosphate a rather large escess of zinc oxide was added, as it was thought that i t would have no effect on the excess of silver nitrate. The results were too high and not in accordance with each other. The excess of the zinc oxide acts upon the silyer nitrate, precipitating i t as the oxide. Experiments have pro\-en this. On allowing 0.3 gram oi zinc oxide to stand in contact with ’5 cc. 0 . 1 -Ysilver nitrate for half an hour, 0.3 cc. of the silver nitrate was found to have been precipitated. Evidently a large excess of zinc oxide cannot be used. The method as finally adopted, 2nd which has given good results, is as fol1on.s : The solution of the phosphoric acid contained in a graduated ( 2 0 0 cc.) ilask is nciitralized, if necessary, to phenolphthalein, with caustic soda free from chloride,’ an escess of 0.1 *V silver nitrate added and the mixture well shaken. Zinc oride, or a suspension of i t in water, is then added little by little, shaking vigorously for two or three minutes after J . SOC.Cheni. Irtd., 1910, 794. If chloride-free soda is not a t hand, it can be neutralized with standard alkali of ktiown chloride content.
I 103
A VOLUMETRIC ESTIMATION OF PHOSPHORIC ACID.
each addition and testing for free acid each time by dropping in a very small piece of blue litmus paper. When the color of the litmus paper is no longer affected, the solution is made tip with water to the mark well mixed and allowed to subside for a fern minutes. It is then filtered through a dry filter and the excess of the silver determined in an aliquot portion of the filtrate ( IOO cc.) by Volhard's method. Some of the results obtained by this method with sodium and ammonitim phosphates and phosphoric acid are here given: I
.............................
P*O: fOUlI(1. Grain,
0.0620
0.0618
-0.0002
0.0622 0.0928 0.0604
--o .0002
............................. 0.0620 3............................. 0.0930 4 ............................. 0.0604 5 ..................... . . . 0.1208 2
O.IZI0
0.0613 0.0916 0.0807 0 . I342 0.2018
....
Error. Gram.
PaOn present. Gram.
0.0692 0.1037
0 . I039
so.0002
+
0.0000
0.0002
0.0000
-0.0003
+
0.0002
--o.OOOI
+0.0003
+
0.0000
0.0003
The points to be observed in this method are : ( I ) A minimum excess of 30 per cent. of the theoretical amount of silver nitrate required must he added, otherwise the results are apt to be too low. A large excess of silver nitrate, even as much as two and a half times the theoretical quantity required, does not interfere. ( 2 ) After the zinc oxide has been added, the mixture should not be allowed to stand long before filtering, as otherwise the excess of the zinc oxide-and an excess of the latter must necessarily be added in order to combine with all the acid-will precipitate some of the silver and thus vitiate the results. It is best to filter as soon as possible. The final volume to whjch the solution is made up has no effect. Using the same quantities of sodium phosphate and silver nitrate, but varying the final yolume from 100--250 cc., no appreciable difference uf results could be obseryed. The method is available in the presence of sulfates, nitrates, chlorides and calcium, but cannot be used for the determination of the total phosphoric acid in the presence of more than traces of iron and a.luminum. !f a chloride is present, a correction must be applied. Since, as has been shown, phosphoric acid is liberated in the interaction of sodium phosphate and d y e r nitrate, the action of the zinc oxide is probably this : it combines with the phosphoric acid, forming zinc phosphate which is then transposed by the silver nitrate into silver phosphate and zinc nitrate. This assumption has led to an investigation of the action of a solution of silver nitrate on the precipitated phosphates. The
1 io4
GENERAL,
rmsxx,
AND ISORGANC.
results may be summarized in the rough generalization that “ precipitated phosphates soluble in acetic acid are transposed, more or less readilj-, by a solution of silver nitrate into sih-er phosphite and the corresponding nitrate, while those insoluble in acetic acid are not so transposed.” Thus, the phosphates 01 the alkaline earths, magnesium, zinc, manganese, nickel, cobalt, copper and cadmium, are transposed, but tliose of iron, aluminum, lead and bismuth are not. Quarititati-,-e examination of the action of sili-er nitrate on precipitated phosphates was liinited to that of calcium, as this is the only phosphate that has any commercial value. It was found that precipitated calcium phosphate is quantitatively transposed into silver phosphate and calcium nitrate when i t is shaken up for a few minutes with a t least twice the theoretical amount of 0.1N silver nitrate required. The procedure is the same as gil-en under soluble phosphates. I n case of the tri-basic phosphate, no zinc oxide is required; as there will be no acid formed. The commercial phosphate is usually a mixture of the di- and tribasic, and some acid will he formed when silver nitrate is added to i t and consequently zinc oxide d l have to be added. Eelow are given some results obtained by this method xith two samples of calcium phosphate. Both contained chloride, which was determined and corrected for. The figures given in the first column were obtained gra-iimetrically. Per cent. P 2 0 j Per cent. P20s present. found.
Error found.
I ..........................
47.20
47.12
-0.08
49.07 48.36 48.34
-.I1
4. . . . . . . . . . . . . . . . . . . . . . . . .
47.20 48.52 48.52
-0.16
-0.18
The fact that the results obtained by this method are slightly lower than the gravimetric is probably due to the presence of a little iron in the phosphate. The operation by the described method should not be performed in direct strong sunlight, as the yellow sil.ver phosphate is liable to blacken somewhat and thus give rise to doubts as to the correctness of the procedure. LABORATORY OF P ~ ~ E R S - W E I ~ H T ~ S A ~ - R O S E C~ o .G , ARTE~ ANALYTICAL D B P A R T X E N T . P H I L A D E L P H I A .
THE OXIDATION O F FERROUS SALTS. BP C H A R L E S BISKERVILLE I X D R E S T O X S T E V E T S O ~ Received April 26, 1911
In the course of an examination of commercial oxygen, i t R ’ ~ Sobserved with surprise that a slightly acid solution of ferrous ammonium sulfate (Mohr’s salt) was not appreciably oxidized by bubbling oxygen through i t for three hours. Subsequent tests with variations were made and this cornmilaication records our findings.
